Building better settings for online content curation

Content curation algorithms today play a large role in deciding what news, social media posts, recommendations, and ads we see online. Major platforms’ designs and deployments of these algorithms, however, have eroded users’ autonomy – their ability to govern for themselves how they want to live their lives. The algorithms often act against user wishes, and promote content that invites user engagement as opposed to high-quality content. Users experience algorithm-promoted misinformation and algorithm-induced filter bubbles. The system of targeted advertising that platforms rely on for profits facilitates mass manipulation. And platforms and their algorithms effectively impose content curation decisions on their users. To give users autonomy, we must empower users to align content curation systems with their values. Control settings – user interface widgets that invite users to adjust and steer content curation algorithms – promise to do just that. But good design of settings remains a challenge. This thesis contributes guidelines and methodology for making more usable and useful settings. First, I present ways that we can make settings easier for users to find and understand. Second, I demonstrate that users’ expectations towards a setting increased their satisfaction towards feed content. Lastly, I show that users don’t want direct control over everything, and suggest factors that increase the likelihood that users will want direct control. These findings help build a framework that can generate design guidelines for content curation settings, with which regulators can eventually take action to ensure platforms deploy content curation systems that support user autonomy.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Computer Science

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2024

Author dc:creator

• Hsu, Silas

Contributors dc:contributor

• Karahalios, Karrie
• Wang, Yang
• Sandvig, Christian
• Cobb, Camille
• Saha, Koustuv

• Autonomy
• Agency
• Control
• User Interface
• Human-computer Interaction
• Advertising
• Choice

Statement dc:rights

• Copyright 2024 Silas Hsu

Language dc:language

en, eng

Handle dc:identifier

https://hdl.handle.net/2142/125687

OAI record oai:identifier

https://www.ideals.illinois.edu/oai-pmh?verb=GetRecord&metadataPrefix=etdms&identifier=oai%3Awww.ideals.illinois.edu%3A2142%2F125687